Estimation of Automobile Demand in the U. S. using random coefficient logit model

点击翻译按钮获取中文摘要

Estimating the demand of differentiated goods and services is essential to understand important problems in industrial organization and marketing such as the evaluation of the value of a brand or of power to control the market by o particular brand. In a durable goods market, the only data available to researcher are often characteristics of the products and their market shares. To compensate this lack of information, researchers have used discrete choice model based on utility maximization behavior of consumers to link characteristics of the products to their marker shares.In this research, we set out to develop a new model by considerably extending the available models, investigate its statistical properties, and apply the model to the U. S. automobile market to obtain a new body of insight into this very important market.The research proceeded in five phases: First estimating the automobile total cost of ownership (hence force, TCO) in the U. S. market; Second investigating the asymptotic behavior of the random-coefficient logit model including consistency and asymptotic normality and asymptotic variance-covariance; Third extending the random-coefficient logit model by combining the demographically categorized purchasing pattern data and again investigating the asymptotic behavior of this extended random-coefficient logit model including consistency and asymptotic and asymptotic variance-covariance; Fourth their applications to the U. S. automobile market using the TCOs rather than the purchasing prices; Five as a precursor to the next phase of research, developing a full information likelihood Bayesian estimation methods.The first three are sequentially completed respectively during 2004-2005, 2005-2006, 2006-2007 academic years, but the last two topics are concurrently dealt in 2007-2008 academic year. We firmly believe this research will contribute to the body of literature in applied industrial organization and marketing.